DeepDNA

DeepDNA gives ChatGPT curated, cited records on pharmacogenomics, so answers about a gene, a variant or a drug–gene pair can quote their sources.

Five read-only tools:

• check_medications: splits a list of up to 30 generic drug names into drugs with a curated drug–gene record and drugs not covered, and says whether each record rests on a CPIC or DPWG guideline, a drug label or the primary literature.
• lookup_drug_gene: the record for a gene, or for one gene and one drug, with the guideline where one exists, any phenotype groups it distinguishes, a summary and sources.
• lookup_gene: everything curated about a gene, including what a consumer genotyping array can and cannot tell you about it.
• lookup_variant: one variant by dbSNP rsID, with its gene, alleles, effect and sources.
• list_genes: what the knowledge base covers.

Coverage is small: 16 genes (CYP2C19, CYP2C9, CYP2D6, CYP3A5, DPYD, TPMT, NUDT15, SLCO1B1, VKORC1, UGT1A1, G6PD, MTHFR, COMT, HFE, APOE, BRCA1) and 83 drug–gene pairs across 15 of them; 72 rest on a CPIC or DPWG guideline. A drug that is not covered is reported as not covered, which does not mean it has no pharmacogenomic interaction.

What it does not do: it takes no genotype or DNA file (the tools have no field for one). Records report what a guideline, a drug label or the primary literature says, which can include dose adjustments by phenotype; the tools do not work out a metabolizer status or a dose for a person, and they give no diagnosis. Answers are written by ChatGPT, an AI model, using these records. Every record comes with the same notice: educational information only, not medical advice. Decisions about a medication stay with the prescriber.

No account and no key. We do not store IP addresses, request bodies or medication lists. Each request records the tool, the gene, rsID or drug looked up (for a medication screen, only how many drugs), the client and a country code; see the privacy policy.